How to classify proteins structures in an objective and unambiguous way is a long-standing problem in protein structure recognitions. We are developing a graph theory approach to give a rigorous parsing of different possible chain folds into conformational taxonomies. Specifically, we are treating secondary structural unit (e.g., a helix) as a vertex on a planar graph. The existing graph theory then gives a rigorous mathematical description of the graph. Such a classification scheme is being used to predict protein structures for given sequences. We use the resources at the Computer Graphics Laboratory to visualize our results.